I am trying to connect to pgsql using ODBC from Visual Objects (for the ones who don't know what is Visual Objects, it's an IDE with its own object-oriented language made by Computer Associates. VO's roots are related to Clipper. ).
The time it takes to get the list of the tables, and then to get the data and display it way too long. It took more than 16 seconds to
connect, get 200 records and print them on the console.
I know for sure that it's not pgsql's fault: with a small c program that uses libpq i got the same data instantaneously. Also, it looks like it's not VO's fault (i know that there are people who used VO against big databases). So, the problem must be related to the ODBC (the pgsql driver or something else?).
Anyway, i've tried to open the same table from MSAccess, and the performance was even worse. Just an example: for going from one record to the next one it lasts about 4-5 seconds.
The machine is a Pentium2 266mhz (128mb) laptop running windows NT.
Today, I have tried to access pgsql from MSAccess at work. It worked very fast. I've used the same ODBC drivers, but the machine is
faster (p4 2.4ghz), has more memory, and the OS is windows 2000.
But it is hard for me to believe that a Pentium2 machine with 128mb RAM
can't handle 200 records !
I begin to think that there is something fishy about the ODBC support installed on the laptop (i have recently "inherited" the laptop from someone else, and i haven't reinstalled the os ).
What do you guys think? what could be done to make things go faster?
I have also two more questions. After installing the pgsql odbc drivers, there are 3 drivers that appear: "PostgreSQL", "POstgreSQL legacy" and "PostgreSQL unicode". What's the difference between "PostgreSQL" and "PostgreSQL lecacy"? Are there any options of the odbc driver that could cause a performance boost?
---------------------------(end of broadcast)--------------------------- TIP 8: explain analyze is your friend